Product Overview โ Mid-Range AI Server for Vision Analytics
The Metrici MidSYS-AI is a balanced, mid-range AI server built to run Metriciโs deep neural network engines in real time. It is designed for growing operations โ sites that have outgrown a single-camera edge device but do not yet need a rack-mounted, multi-GPU platform.
The unit is delivered as a ready-to-use appliance: the operating system and Metrici engines are pre-installed, so no software setup is required on site. The engines are then customised for the specific deployment โ adding cameras, configuring detection parameters, and completing setup in the Metrici Web Interface. All processing happens locally on the unit; no data has to leave your premises.
Metrici AI Servers come in four tiers, from a compact edge device to a data-center-scale platform. MidSYS-AI sits in the mid-range โ the natural step up for small business, retail, and lighter industrial deployments.
Each tier is pre-validated with Metrici software before delivery, and the architecture is designed to scale โ you can move up to a higher tier as a deployment grows. If you are unsure which model fits, Metriciโs team sizes the system based on your cameras, engines, and expected workload.
The full specification of the MidSYS-AI as delivered:
| Specification | Details |
|---|---|
| Architecture | Intel x86_64 |
| Processor (CPU) | 1 × Intelยฎ Coreโข Ultra 5 245K, 4.2 GHz |
| Processing units | 14 cores (6 P-cores + 8 E-cores) / 14 threads |
| AI accelerator | 1 × Intelยฎ AI Boost NPU (integrated), up to 13 TOPS (INT8) |
| Performance | Up to 8 cameras, depending on workload type |
| RAM memory | 16 GB DDR5 |
| Storage | 500 GB SSD |
| Networking | 1 × Ethernet 10 / 100 / 1000 Mbps |
| Graphics | Integrated Intelยฎ Graphics (4 Xe-cores), HDMI output |
| Format | Cube case, 34 × 32 × 43 cm |
| Power source | 1 × 850 W |
| Operating system | Linux Ubuntu 24.04 LTS |
The MidSYS-AI is configured to run any combination of Metrici detection engines at the same time. Common engines include:
The unit supports up to 8 cameras. The exact number it can drive comfortably depends on the engine mix, frame rates, and detection load โ a single light engine reaches the upper figure more easily than several heavy engines running together.
The MidSYS-AI arrives ready to use. To bring it online for the first time:
The operating system completes boot within roughly 30โ60 seconds, and the Metrici engines start automatically as background services. The unitโs IP address is either pre-configured by Metrici before shipping (documented in the delivery note) or obtained from your network by DHCP. From there, configuration is completed in the Metrici Web Interface.
Install the unit on a stable surface in a well-ventilated area, away from direct sunlight, heat sources, and high-dust environments. Leave clearance around the chassis so air can move freely through the vents, and avoid enclosed cabinets without active ventilation.
For configuration help, system sizing, or a quote for the right server tier, contact the Metrici team. The product datasheet and CE declaration are linked in the sidebar under Resources.